Estimating Quadratic Almost Ideal Demand Systems (QUAIDS)

[Kiplangat Cheruiyot]

Dear Statalists,
I want to estimate household energy demand using cross-sectional survey data. My question is how to compute respective energy prices and budget shares so that I can make use of the aidsills command or Poi's (2012) quaids command.
I have attempted to create respective energy prices and budget shares. In trying to estimate Quaids model proposed by Banks, Blundell and Lewbell I get an error message "No observations" What can be the ready to this?

[David Achiles]

about how to compute prices and budget share, it depends on you data, basically most of households microdata do not present prices, therefore, idea coming up to compute unite values (expenditure of each products divided by quantity consume).

It would be very useful if you could present some detail information about you data

[Kiplangat Cheruiyot]

I am using Integrated household budget survey data. Households use only one of the energy options (Firewood, Charcoal, Kerosene, LPG or Electricity) and not a combination. I therefore would like to compute the budget shares relating to Firewood, Charcoal, Kerosene, LPG and Electricity as well as their respective prices. The data available relate to energy expenditure and the units consumed of each (which makes it a bit easier to compute prices).
I wish to obtain budget shares of Firewood, Charcoal, Kerosene, LPG and Electricity which ensuring that the adding-up restriction imposed by theory is not violated.

I look forward to your responses.

Thanks!

[Jorge Eduardo Perez Perez]

To calculate the budget shares, you just need to divide expenditure in each category over total expenditure. You may want these as shares of total expenditure, or as shares of energy expenditure, depending on what you want to estimate. For example, if firewood is expenditure on Firewood and energy_expenditure and total_expenditure are the energy and total expenditure variables.

gen share_firewood_energy = firewood/energy_expenditure
gen share_firewood_total = firewwod/total_expenditure

You can not obtain prices in this dataset, just unit values. There is a subtle difference. You should look at Deaton's "The Analysis of Household Surveys" for a discussion of unit values. To obtain them, just divide the expenditure in each category over the number of units consumed.